I do not think this is expected – if it is – than need to know why. I did this because a friend reported he had an issue last night. Seem I have found the same thing he did.
Did a fresh upload from ‘new’ RM Database to a new Ancestry tree – counts do not match (1120 diff). Double checked in SQL
-- RM not in Ancestry QTY
SELECT Count(PersonID) as QTY
FROM PersonTable
LEFT JOIN AncestryTable ON PersonID = rmID
WHERE anID is null
count was 1023 diff.
Looked at who they were.
-- missing Ancestry people in RM by PID
SELECT PersonID as PID, ParentID, SpouseID, Relate1, Relate2, Flags, Living, anID
FROM PersonTable
LEFT JOIN AncestryTable ON PersonID = rmID
WHERE anID is null
It is puzzling as the missing people (at least first glance seem to be random) - RM id all over-- different relationships to set relationship person (or root for that matter)
Do a Find Everywhere search of the database for “class.”, without quotes and see if anyone has it. We know a batch will not upload if what appears to be a code to Ancestry is in there. They will need to rearrange the sentence in RM so class does not end with a period.
I just now disconnected my RM10 database from Ancestry and deleted the tree from Ancestry. I do this periodically anyway, and I had not done so since I quit running RM7. The tree in Ancestry still had the RM7 tree name, but RM10 accessed it just fine because I obviously had imported from RM7 to RM10 to get the RM10 database started. I then uploaded my RM10 database to a new Ancestry tree. The upload did not report any errors,
After the upload, my PersonTable in RM has 39545 rows and my AncestryTable in RM has 35763 rows. That’s a pretty substantial discrepancy. I tried Renee’s Find Everywhere for class. There were some hits, but the number of hits was a couple of dozen, not thousands.
These days, I access Ancestry data from RM mostly by Right Click > Jump To rather than by looking at hints in RM. That still works except for the 4,000 or so people who didn’t make it from RM to Ancestry. I may pursue the problem further at some point in the future, but I would rather work on other things right now, I do not see a pattern in the people who were not uploaded, but I really didn’t look very hard for such patterns.
that does not appear to be the issue – many people have not facts or sentences – so it is something else. Not obvious patterns some related, some not, some with spouses or parents(or not), some with FamSearch ID (check that for just in case). Some have No notes or sentences.
okay I have discovered something that I can not explain as I do know why this should occur.
I have my friend send me 2 of his databases I was able to upload to ancestry under my account.
So the key thing might be the ancestry account itself (and maybe the link IDs or something on the ancestry side) . Note- I did disconnect…
If you have “class.” in the batch being sent it won’t upload. Not sure how many people get sent at a time but if it were 100 people and only 1 person has class ending with a period the entire batch is rejected by Ancestry. If you have class in the middle of a sentence it’s not a problem. It has to have the period at the end of that word to look like a code. The use of class. could be anywhere in the file connected to data being sent. The most likely place is notes or sources.
respectfully based on my test and observation that statement is 100% false.
It may be reason while it fails sometimes. But that would not explain why the exact same database failed on one Ancestry acount but not the other. And that one had “case” in notes. You need to explain how one upload can fail but not the other (again same database but different account) has 100% of all people.
Now the batches would explain why /where people appeared to be random over whole database when it did fail.
Why one account would accept the upload of a database and not another would be something Ancestry would have to look at. Are you logging in on the same computer?
Since its going through RM software that would be something RM need to work out with Ancestry. (not me).
Different computers and locations in USA — I suppose it could possibly related to being connected to different severs. Bottom-line there is some beyond “race” and periods – if results are not consist when Race and periods are include and work sometimes then it may not even have to no with either and/or both. I used to work at a college where I troubleshooted what systems were contributed the issue (root cause etc) that was about 10 years --but I identified the issue 99% of the time.The evidence I found contradicts your statement ‘will cause faiulre’ (vs may cause failure). Not sure it possible yet to find the true contrubuting factors.
The key evidence I found since the “tree people”(of RM & ANC database) and former connections were were not on the opposite account – that seems to be a more likely contributing factor – but unfortunately my ability to test that is rather limited
Ok. I have remained silent but cannot hold back any longer. I am the friend that kevync1985 has been referring to. He and I have spent countless hours determining if certain issues are caused by us or are issues with the application. Both he and I are experts working with data and I am also an application developer. You made one statement which should have stopped all of this conversation:
“We know a batch will not upload if what appears to be a code to Ancestry is in there.”
Instead of blaming or telling your customers how to work around a bug, FIX IT! Don’t tell me I need to spend time combing through my data looking for something that will cause a problem with your code. A good application developer will always make sure that a user cannot cause an issue with their coding. A simple error routine with a message would go a long way to avoid people wasting time.
Besides, like kevync1985 said, the second time I tried all of my ancestors got copied to Ancestry.com. I wasted four hours trying to figure out why 50 of my people did not get copied when I created and uploaded my RM database to Ancestry.com. I also involved my friend who posted for me.
By the way, I have had a tree on Ancestry.com for a very long time. Recently I have been having a lot of weird issues with synching that I was trying to figure out. Then I read a post in which you explain something to the effect that a tree created in RM 7 my have issues after updating to RM 10. Thanks for letting your users know that when you released version 10!
I don’t like posting issues because people always try to tell me that I must be doing something wrong. You just proved my point.
Summary:
any time try to upload and create new tree on Ancestry I first disconnect in RM – it is possible it momentarily connects before I disconnect.
Not sure if API / server issue – but its not simply because of ‘class’ or period. Otherwise at least one group would fail and that does not happen consistently.
Database
Computer
Account
Comment
Result
A
A
A
Has ‘class’ in notes and period in some sentences
Failed (about 4% people missing)
A
B
B
Same above with class and period - diff account and people and/or similar tree did not exist previously not sure if coincidence on that part
100 % success tried same day
B
B
B
Has period in some sentences and ‘classes’ but no ‘class’
Failed (about 5% of people missing)
B
C
B
(same as above)
Same DB to new diff Anc tree no failures – 100% success
B
A
A
(same as above)
Not let tested
B
B
B
Removed all other existing trees on anc & Disconnected RM then uploaded to new Anc tree name (ending in v2)
Same DB to new diff Anc tree no failures – 100% success
RM has no control over what Ancestry sees as code. There is nothing for us to fix on our end. How do you stop people from writing sentences that end with “class.”? If we somehow stripped it out wouldn’t you be just as upset to see it missing? It is extremely simple to search for the use of “class.” in the entire database using Find Everywhere. You can even rewrite the sentence when reviewing the results. I highly doubt Ancestry will remove their restriction because allowing coding in could potentially take their website down. Which is also the reason they will no longer accept the upload of HTML files.
Open a support ticket with a backup with media included. We will test if we can upload it. If we can then it’s not RM or the database, but something on the computer or server you are hitting. Tell the support agent if the file is to large to upload in the ticket. They will send you a link to upload it. https://support.rootsmagic.com/hc/en-us/requests/new
Are you using speech-to-text? So many of your sentences sound strange and it’s hard to make sense of them. Renee has warned about “class.” but you have been responding about “case” and "race’. Is that just a mistranscription? It would help readers if you proofread before posting.
The table of tests was most interesting. Results seem almost random. Too many variables seem to be in play.
If a RM database has multiple trees, do all of them upload to the Ancestry Tree?
I am not going to keep going back and forth. My point is there are known issues and your responses have been “Yep, we know about that and here is what YOU have to do”. I don’t like running for support every time I have an issue but that is what you are asking your users to do. Instead of spending time seeing if I did something wrong, I should have just posted immediately “Tried to create and upload Roots Magic database to Ancestry.com but not all of my users got uploaded.”
If this is an issue from Ancestry, then take some responsibility for customers and create a notice about this.
Here is the bigger problem. I don’t have any use of the word “class.” and everything worked fine the 2nd time. Again, I should have received some type of error on the first run and you haven’t explained why it worked the 2nd time.
You have not determined the point of failure(s) in my examples.(Maybe ‘class.’ at end may cause error/failure – I do not have a example to test that). Not sure if related to API or server – while I agree the point of failure might seem to lean/point to Ancestry not enough info to blame Ancestry. As for the issue with a specific word — I have seen very similar issue when I was with the college. Which the software had to fix (in next major update).
Here is what we know:
Same dataset when fails 1st time will pass after/when several attempts taken – at first I thought it might have been related to account but that does not seem to be the case
other than any minor change in database during close and open RM might make - its is the exact same database
one database as about 2.3K people (A) and the other about 25K people (B)
